Combining Strategies/ Sizing

[minx]

I'm working on 3 systems which are closely related to each other but have different exit points. I'm going to allocate 1/3 of my capital to each and increase my size as profits are made and vice versa. The argument I'm having is whether to use the percentage increase in combined capital to calculate the size for each strategy or whether to use their individual capital increase.
The advantage of the first is that the portfolio remains balanced throughout its life, I'm concerned that one system may run away from the others and lead to one of them being a 'runt' system. The disadvantage is that individual systems may experience worse drawdowns than they should had they been operating independently.
I can imagine that some of you would just say that I drop one of them if they become a 'runt' system but they each serve a purpose in different market conditions and that would actually unbalance the portfolio should mkt conditions change in favour of the 'runt'.
Thanks!

 

[Mr. Charts]

It's not clear whether you are investing and/or trading or something inbetween.
Perhaps it's worth identifying the market conditions obtaining at the time and applying the appropriate strategy. If you cannot identify those conditions, maybe that requires more attention.
Have you back tested throughout different market types?
Have you then forward walked the strategies?
Have you paper traded them first?
IMHO, it is unwise to risk capital to find out what will happen before you have a tested and successful methodology.
Generally, as in all investments and trading approaches, it is better to kill the weeds as early as they are recognisable,(long before they take over), and feed the flowers. Plants which do not thrive in all conditions should be planted only in the correct conditions.
Richard

 

[minx]

I'm trading futures and I have backtested the strategies and traded a derivative of them every day since the beginning of the year.
I'm happy with the results it's giving and the theoretical results the backtesting has produced but now moving forwards I'm interested in how to manage the equity and the relationship between all 3 systems. I'm worried that if the 'runt' falls too far behind then when its moment of glory comes it'll be too small to protect the account from the other 2 systems, also if I use total capital to work out the sizings then in the bad times the runt will keep the size up on the 2 losing systems and in good times the other 2 will keep the runt size high thereby affecting performance.
Just wondered if anyone can shed some light on this, whether its is best to balance the portfolio or whether to let the winners run unconstrained. I have also thought of setting collars so it cant fall below a certain value of the portfolio and cant rise too high either.

 

[Tuffty]

What I've done in similar circumstances is to run the systems independently and then do portfolio re-balancing on a periodic basis (i.e. a time period that you're comfortable with).

The same sort of problem occures when running several systems/or investments that aren't highly correlated. Should you let your winners run, or cut them and re-balance? My conclusion is that you should re-balance every so often as eventually you'll end up with just one outsized system/investment and have no benefit from portfolio theory (i.e. diversification). If you don't rebalance you may as well just have decided to run the best system with all your money from the off.

If you want diversification then you need to rebalance at regular intervals, if you don't want diversification then just run the 'best' system.

 

[Tuffty]

minx, I just remembered what I found when testing this. If you don't rebalance your returns can end up being higher overall overtime but the standard deviation of them from one period to the next is bigger. Horses for courses etc.

 

[jmreeve]

I think I favour a similar approach to Tufty.

Allocate capital to each system based on your total capitalisation and then periodically e.g quarterly or anually , re-optimise the portfolio weightings. This allows the portfolio to grow without becoming unbalanced.

However, it would be better to have more that 3 systems.

 

[commanderco]

Hola Minx,

I agree with TUFFTY.
Keep your systems independant, otherwise you are assuming a co relation
where one may not exist.
If your 3 systems have a very low co relation, as they should have, and not differing
lengths of MA´s etc, then 1 of them at least should be firing.
If your win/loss ratios are robust then you should be covered.

 

[minx]

Cheers all, thing is that they are highly correlated, all have the same entry/ stop triggers. The difference is the exits. I guess you 'could' say its one system with 3 exits, except there is a possibility of a re-entry on some whilst the others dont have this facility.

 

[rdstagg]

Hi Minx

If the conditions for entry are the same but the exits differ you have one system but 3 strategies. This is different from 3 systems that are correlated.

I would export all the trades and then backetest different position sizing strategies on your portfolio.

You might want to read some books
Stridsman,Thomas (all of them)
Jones,Ryan - the trading game (fixed ratio pos sizing)
Van Tharp (all of them)

or buy some cheap software that will do it for you
eg adaptrade.com

However I would read the books first. You MUST understand what you are doing here.
Execute a good plan rather than execute your account!

Hope that helps.

PS It may help if you tell us what your strategies are - who knows someone might come up with an improvement or an observation that could really help?
I recommend looking at
http://www.trade2win.com/knowledge/articles/general articles/developing-a-trading-strategy-2

 

[rdstagg]

Hi Minx

If the conditions for entry are the same but the exits differ you have one system but 3 strategies. This is different from 3 systems that are correlated.

I would export all the trades and then backetest different position sizing strategies on your portfolio.

You might want to read some books
Stridsman,Thomas (all of them)
Jones,Ryan - the trading game (fixed ratio pos sizing)
Van Tharp (all of them)

or buy some cheap software that will do it for you
eg adaptrade.com

However I would read the books first. You MUST understand what you are doing here.
Execute a good plan rather than execute your account!

Hope that helps.

PS It may help if you tell us what your strategies are - who knows someone might come up with an improvement or an observation that could really help?
I recommend looking at
http://www.trade2win.com/knowledge/articles/general articles/developing-a-trading-strategy-2